What we don't know about social anxiety

Why this page exists

Social anxiety is very treatable. Cognitive therapy for social anxiety, internet CBT and SSRIs all have strong trials behind them. That should not hide the gaps. Knowing where the evidence stops helps you judge new claims, and helps you see that “no evidence” often means “nobody has checked”, not “it doesn’t work”.

1. The years before anyone gets help

This is the biggest gap. Across surveys in 24 countries, only about 23 in 100 people with social anxiety ever got treatment. On average, people waited more than 15 years from when it started to their first treatment. A large US survey found delays of 9 to 23 years for anxiety problems.

Those years are almost unstudied. That is when people try self-help books, apps, Reddit, alcohol and chatbots. It is where most readers of this site are. We know very little about what helps during that time, or what makes things worse. If you are in that window, start here.

2. Results that exist but are hidden

Of 471 finished social anxiety trials on the main US registry, only 85 had posted results on 9 November 2026. Some of the missing ones test CBD, MDMA and the nasal spray fasedienol. No new study is needed to fill this gap. The data are already there. See trials waiting for results.

3. Home tech against proven self-help

A review of 21 studies of self-guided VR found no study that compared it with internet CBT or a self-help book. For apps, most trials had a high risk of bias, and the only larger well-run trial found no clear benefit. So we know home VR and some apps beat doing nothing. We don’t know whether they beat the free or cheap options that already work. See VR at home and apps.

4. Whether VR is better than real-life practice

VR exposure inside therapy works about as well as real-life practice. None of the trials in that review was designed to show that the two are truly equal, and none has shown VR is better. We also don’t know how long gains last beyond a year. The VR guide explains what we do know.

5. Chatbots and social VR: practice or avoidance?

Many people now practise talking with AI chatbots or meet people in social VR. Nobody has measured whether this works like practice, which would help, or like avoidance, which would keep the fear going.

• Social VR: users report feeling less anxious inside it than offline. That is equally consistent with practice and with hiding.
• Chatbots: the best trial of an AI therapy bot did not include people with social anxiety.

The question could be answered now, for example by tracking whether people who use these tools see more people offline over time. See social VR and AI chatbots.

6. Teenagers and VR

Social anxiety usually starts in the teenage years. Yet we found no finished randomized trial of VR exposure for teenagers with social anxiety. A trial in 12 to 16 year olds has started. For other treatments in young people, see therapy for children and teens.

7. Older adults

A Canadian survey of 12,792 people aged 55 and over found about 5 in 100 had had social anxiety at some point in their life. Treatment trials almost never focus on older adults. Social anxiety is often seen as a young person’s problem, so older people may be missed.

8. People outside rich Western countries

A survey of 142,405 people in 28 surveys found social anxiety was reported least in low-income countries, and treatment was rarest there too. Some of this may be real. Some may be because the questions do not fit every culture. In Japan and Korea, for example, a form called taijin kyofusho centres on fear of offending others, not only of being embarrassed. Standard questionnaires may miss it. Very few treatment trials have been run in Africa or South Asia.

9. Autistic people

Anxiety is common in autistic adults. A review found that 27 in 100 had a current anxiety disorder and 42 in 100 had one at some point in life. We don’t know whether standard therapy for social anxiety, which focuses on fear of being judged, fits autistic social anxiety. The one MDMA pilot in autistic adults had 12 people. See autism and social anxiety.

10. Work

Social anxiety often blocks jobs and careers. In a US trial of 250 job seekers, adding work-focused CBT to job services eased social anxiety a little at first. The difference faded later, and hours worked did not change. We know almost nothing about what helps people keep a job, or about workplace adjustments for social anxiety.

How to use these gaps

• “Not tested” is not “doesn’t work”. It means you are trying something without a map. That can be fine for low-risk things, like social VR. Watch what happens to your real life.
• Be wary of anyone who fills a gap with certainty. If a product claims to be proven for social anxiety, ask for the trial.
• Start with what is known. The approaches with strong evidence are on what helps.